Influenza (and other (-)ssRNA viruses) are transcribed and replicated by an RNA-dependent RNA polymerase, that is, the protein makes +RNA out of a -RNA template. This replication is happening in the cytoplasm, and for at least a limited amount of time, there is dsRNA present, although I am not sure of the on/off time for this. The paper you cite uses an antibody to detect dsRNA, which may not be an adequate marker for dsRNA ( the RNA nay be in an incompatible confirmation for the Ab, proximety to proteins may cause the RNA to be fixed to viral or other proteins during fixation procedure, etc).